A defined alignment score
A 0–100 photo symmetry score calculated from the average normalized pair difference.
FREE TOOL / YOUR FEATURES IN FOCUS
Face symmetry test
Compare your left and right sides. See paired landmarks, regional alignment, and a symmetry score from one front-facing photo.
Upload a photo
Drag and drop or choose a file
Your first result is free. No account needed.
01A defined alignment score
02All twelve pair differences
03The axis and capture context
SEE WHAT YOU GET
Face symmetry test: a result you can explore.
Inspect left–right differences instead of guessing from a mirror.
See the report in actionFace symmetry: find the difference you can actually see
LIVE EXAMPLE · AI-CREATED ADULT PORTRAIT
See paired positions on the same photograph.
Loading the measured example…
Pairs are reflected around the fitted facial axis. Smaller differences mean closer alignment in this image.
Is the difference in my features or in this photo?
Inspect twelve reflected landmark pairs across the face. Check pose and expression before treating a small mismatch as a physical change.
- 01
Locate the pair
Follow eye, mouth and jaw positions back to the original photo.
- 02
Check the setup
Look for head turn, uneven expression or an obscured endpoint.
- 03
Repeat fairly
Retake at the same distance in even light before interpreting a difference.
WHAT YOU RECEIVE
Face symmetry test results: the paired measurements
All 12 pair differences are open in this test, along with the overall symmetry reading and head-roll context. Inspect the pairs rather than relying only on the summary number.
All twelve pair differences
Open eye, brow, nose, mouth, cheek and jaw comparisons, expressed as a percentage of cheek width.
The axis and capture context
Inspect the paired points, fitted midline and head-roll correction on your original photograph.
Mirror images, camera photos and an asymmetrical face
Mirroring changes which image you see. A symmetry measurement instead defines a comparison within the original photograph; the two experiences need different interpretations.
Flipping changes sides, not the underlying geometry
A horizontal flip swaps which side you call image-left and image-right. With the same coordinates reflected consistently, the paired alignment calculation should remain the same. A new scan may vary slightly because landmark detection is repeated on the flipped pixels.
A mirrored composite answers a different question
A left-half reflection creates a face built from one side, and a right-half reflection creates another. Those images can look striking, but they are constructed pictures rather than measurements of your original whole face. FaceRater’s symmetry result stays anchored to the original paired landmarks.
One number can conceal different patterns
Two photos can share a score while having different pair-level differences. Read the mouth and eyelids alongside expression, then compare the cheek and jaw pairs. For an outline question rather than alignment, use the separate shape detector.
Check your visible face shape ↗Face symmetry test: how it works
Choose the question you want to explore. The definitions, examples and research are here when you need them.
Facial symmetry tests and mirror filters answer different questions
A mirrored image can be striking, but it changes the image you are viewing. A paired measurement keeps the original photo and describes selected differences within it.
| View | What happens | What to conclude |
|---|---|---|
| Ordinary mirror view | The whole image is reversed left to right. | Orientation changes; no new face measurement is created. |
| Left-left or right-right composite | One half is duplicated to construct a new face. | Useful as an image experiment, not the original person. |
| This paired-landmark test | Selected points are reflected about a fitted axis and compared. | A defined geometric difference within the submitted photograph. |
What does a facial symmetry test measure?
This face symmetry test compares selected left and right landmark positions around a fitted axis. It keeps the original photograph visible so you can inspect the source of a difference.
A comparison across a declared midline
This test compares twelve pairs of predicted feature points across a fitted axis. It asks how closely reflected coordinates line up in this image. It does not compare skin texture, diagnose a facial condition, or measure symmetry of the underlying skeleton.
Each pair has its own normalized difference
We reflect one point across the axis and calculate the distance to its partner. Dividing that distance by visible cheek width makes the result a relative reading. A 2% width difference means two hundredths of the cheek span, not two millimetres or 2% less attractiveness.
Small differences are not a diagnosis
The face is not a manufactured object with a required perfect match. The clinical literature discusses multiple ways to assess asymmetry. Our single-photo tool provides descriptive geometry, with no validated cutoff separating a normal face from an abnormal one.
How the symmetry score and twelve readings are calculated
The calculation uses reflection and a cheek-width normalization. The summary and the individual pair differences have different units and should be read separately.
The exact score transformation
In geometry-v1.1, the score is 100 minus seven times the mean of the twelve pair differences, clipped to 0–100 and rounded. For example, a mean difference of 2% produces 86/100. That multiplier is our product choice; the score is not a clinical index or population percentile.
Which features are included
The twelve pairs cover outer and inner eye corners, upper and lower eyelids, outer brows, brow peaks, inner brows, nose wings, mouth corners, cheek contours, jaw contours and lower jaw points. All twelve are open in the free result, so you can inspect what drives the average.
What a connecting line means
A line joins the detected pair to help you find it. Its length is not the asymmetry error: most of that length is simply the distance across the face. The reading instead uses the reflected discrepancy relative to the fitted axis and cheek-width normalization.
Review the coordinate and score definitions ↗How to check your facial symmetry
FROM PHOTO TO REPORT
Keep a relaxed front view, inspect the paired differences and check any surprising reading on the photo. Saving the result is optional.
- 01
Face the camera directly
Use an eye-level portrait with a relaxed mouth and both eyes visible. Keep your head upright without turning toward one shoulder. Confirm adult photo permission and choose the image.
- 02
Inspect the fitted axis
Open the result and check the points against your features. The brow-area-to-chin axis corrects image-plane rotation. It does not reconstruct the face in three dimensions or correct a turned head.
- 03
Open each paired reading
Compare the eyes, brows, mouth and jaw individually. Select a pair to highlight it on the photo. If a large difference follows a pose or expression change, take a matched portrait before interpreting it.
Use a facial symmetry test to investigate the photo
A USEFUL NEXT STEP
The useful next step after an unexpected result is to locate which pair changed, then check how the portrait was taken.
Check a difference around the eyes
Open the pair and inspect visibility, expression and head turn. Reflections on glasses or partly closed lids can alter the predicted points. Use a clearer relaxed portrait before interpreting a small difference.
Check the mouth after a smile
A smile moves the mouth corners and nearby soft tissue. Compare relaxed images if the question is repeatable alignment. Do not read an expressive photograph as a stable map of the face at rest.
Check one apparently wider jaw side
Read the jaw pairs with the contour overlay and head pose. Turning slightly can change the visible width. Keep both sides visible and compare a matched capture instead of applying a mirror filter as proof.
Facial asymmetry research and single-photo measurements
Clinical literature and a browser photo test work at different levels. This tool describes selected image geometry and provides no diagnosis.
Assessment is broader than one summary score
Choi’s review describes multiple approaches to facial asymmetry assessment. That breadth is a reason to keep this single-image result narrowly defined: it does not establish the cause or clinical significance of a visible difference.
An image-plane correction has a limited job
FaceRater removes fitted head roll before pair comparisons. Rotation correction aligns the photo’s coordinate system; it cannot restore hidden depth or undo the perspective caused by turning toward or away from the camera.
Make the result inspectable
Read a particular pair’s percentage alongside its plotted points. A percentage of cheek width is a normalization choice, not a percentage of “facial imperfection”. It lets the report state its calculation without assigning a social meaning to the difference.
Make your face symmetry check more repeatable
A consistent portrait makes repeated measurements easier to understand. Check roll, turn, expression and visibility before interpreting a small difference as stable alignment.
Head roll is only one part of pose
The calculation removes rotation within the image plane. Turning left or right changes apparent widths; looking up or down changes visible relationships. An upright-looking image can still have perspective effects. A retake is more useful than trusting a correction to solve every pose issue.
Use a relaxed expression and even light
Smiling, squinting or raising one brow moves the points being compared. Hair and glare can obscure a feature. Choose a clear baseline and check the overlay before using a higher pair difference as a reason to investigate that region.
Compare the same readings under the same conditions
Keep distance, framing and expression similar. A changed score is not proof of a structural change. If your interest is specifically eye-corner direction, the tilt tool provides two angular readings rather than combining eye position into a whole-face average.
Measure left and right canthal tilt ↗Sources: Ferrario et al. (1993) · Craniofacial morphometry by photographic evaluations
BEFORE AND AFTER YOUR SCAN
Questions about the face symmetry test
These answers cover the face symmetry test questions that affect how you use and interpret the result. Open a question for its measurement definition or a practical next step.
Is this facial symmetry test free?
Yes. The alignment score, all twelve pair differences, head-roll reading and overlay are free. The optional full report adds other geometry readings, regional explanations, a plan and PDF.
What is a good face symmetry score?
There is no validated universal pass mark for this product formula. A higher value means closer predicted alignment in this photograph. It does not establish attractiveness, health or a population rank.
Does an 86 symmetry score mean my face is 86% symmetrical?
No. It is a transformed alignment score. With this formula, a mean pair difference of 2% of cheek width gives 86/100. The score is not the percentage of your face that is symmetrical.
Can it tell me why my face is asymmetrical?
No. It describes predicted coordinates in a photograph. It cannot determine an underlying cause or distinguish skeletal differences from soft tissue, pose and capture effects.
Why do my mirror and selfie look different?
A mirror reverses left and right, while camera distance, orientation and expression can also change the image. Compare the same photograph and its flip separately from a newly captured selfie.
Can I inspect individual eye and jaw symmetry for free?
Yes. Select the relevant paired reading in the result to highlight its landmarks. The reported difference uses the fitted midline and is normalized to visible cheek width.
Face symmetry test: sources and measurement notes
Written and reviewed by FaceRater · . Research provides context; it does not validate our scoring formula. Product definitions and limitations are explained in the measurement method.
- Choi (2015) · Analysis of Facial Asymmetry
Clinical review of asymmetry assessment; does not validate this single-photo score or provide its thresholds.
- Ferrario et al. (1993) · Craniofacial morphometry by photographic evaluations
Primary photographic measurement study motivating consistent capture. Clinical protocols and physical dimensions are not reproduced by this browser tool.
THE COMPLETE FACE REPORT
See the complete sample Want to connect symmetry with the rest of the portrait?
The pair measurements stay free. The complete report adds the other width, proportion and expression readings, so you can read a difference within five regional chapters and keep the explanation with a practical plan.
Try the face symmetry test.
Inspect left–right differences instead of guessing from a mirror.
Explore the sample report